Development k3d cluster automation📜
NOTE: This script does not does not install Flux or deploy Big Bang. You must handle those deployments after your k3d dev cluster is ready.
The instance will automatically terminate 8 hours after creation.
Install and Configure Dependencies📜
- Install aws cli
curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
# sudo apt install unzip -y
unzip awscliv2.zip
sudo ./aws/install
rm -rf aws
rm awscliv2.zip
aws --version
- Configure aws cli
aws configure
# aws_access_key_id - The AWS access key part of your credentials
# aws_secret_access_key - The AWS secret access key part of your credentials
# region - us-gov-west-1
# output - json
# Verify configuration
aws configure list
-
Install jq
Follow jq installation instructions for your workstation operating system.
https://stedolan.github.io/jq/download/ -
Mac users will need to install the GNU version of the sed command.
https://medium.com/@bramblexu/install-gnu-sed-on-mac-os-and-set-it-as-default-7c17ef1b8f64
Usage📜
The default with no options specified is to use the EC2 public IP for the k3d cluster and the security group.
./docs/assets/scripts/developer/k3d-dev.sh -h
AWS User Name: your.name
Usage:
k3d-dev.sh -b -p -m -a -d -h
-b use BIG M5 instance. Default is m5a.4xlarge
-p use private IP for security group and k3d cluster
-m create k3d cluster with metalLB
-a attach secondary Public IP (overrides -p and -m flags)
-d destroy related AWS resources
-w install the weave CNI instead of the default flannel CNI
-h output help
To use a different AWS profile, VPC, or AMI📜
The script uses the default AWS profile and defaults the VPC id and AMI id. To change this default behavior you can export environment variables before running the script.
Run the script with a specific AWS profile by first exporting the AWS_PROFILE environment variable:
export AWS_PROFILE=my-aws-profile
export VPC_ID=vpc-XXXXXXXXXXXX
export AMI_ID=ami-XXXXXXXXXX
export K3S_IMAGE_TAG=sometag
You can also run these inline with the script, without exporting them to your environment. Example:
K3S_IMAGE_TAG=v1.24.12-k3s1 ./docs/assets/scripts/developer/k3d-dev.sh
After Running The Script📜
Follow the instructions from the script output to access and use the cluster.
Install FluxCD📜
The Big Bang product is tightly coupled with the GitOps tool FluxCD. Before you can deploy Big Bang you must deploy FluxCD on your k8s cluster. To guarantee that you are using the version of FluxCD that is compatible with the version of Big Bang that you are deploying use the Big Bang provided script. You will need your Iron Bank pull credentials and command line access to the k8s cluster from your workstation.
./scripts/install_flux.sh -u your-user-name -p your-password
Deploy Bigbang📜
From the bigbang directory deploy BigBang via helm
helm upgrade -i bigbang chart/ -n bigbang --create-namespace --set registryCredentials.username=XXXXX --set registryCredentials.password='XXXXX' -f chart/ingress-certs.yaml -f chart/values.yaml
Overrides can be supplemented by adding references to the specific yaml file, the right-most values file will take highest precedence:
-f ../other-overrides.yaml
Testing Keycloak📜
Refer to this documentation for various options for testing Keycloak which requires two ingresses on the same EC2 instance.
Troubleshooting📜
- If you are on a Mac insure that you have GNU sed command installed. Otherwise you will see this error and the kubeconfig will not be updated with the IP from the instance.
copy kubeconfig
config 100% 3019 72.9KB/s 00:00
sed: 1: "...": extra characters at the end of p command
-
If you get a failure from the script study and correct the error. Then run script with “-d” option to clean up resources. Then re-run your original command.
-
Occasionally a ssh command will fail because of connection problems. If this happens the script will fail with “unexpected EOF”. Simply try again. Run the script with
-dto clean up resources. Then re-run your original command.